python - Pandas - Create new dataframes based on ranked values in select columns -

python - Pandas - Create new dataframes based on ranked values in select columns -

i have dataframe columns containing numerical info , other containing text. looks like:

age weight blood sugar study grouping gender notes 29 195 126 b female notes of kind 34 180 140 b male different set of notes 48 220 111 c male blah blah 55 189 109 c male more notes

i want create sub-divisions of info frame based on rankings of numerical info columns. example, if need 2 oldest patients new dataframe this:

age weight blood sugar study grouping gender notes 48 220 111 c male blah blah 55 189 109 c male more notes

the rank function looks useful. figure run:

df2 = rank.df(axis=0)

and find way utilize index of df2 pull rows df new dataframes. along lines of:

cutoff = df2[df2 > 10] # delete rows nan values in columns of involvement

this feels bit clunky though. i'm hoping there's more straight-forward way say,

"pandas, want new dataframe 15 oldest people in one. great! want new dataframe 20 youngest people, etc"

one alternative sort dataframe age:

df = df.sort('age')

then age of n-th youngest person df['age'].values[n] , age of n-th oldest person df['age'].values[-n].

therefore, view dataframe people of 15 oldest ages, do:

df[df['age'] >= df['age'].values[-15]]

alternatively, if want limit number of rows returned (e.g. don't mind there may 20 people sharing oldest age of, say, 55), utilize head , tail methods on sorted dataframe...

df_age = df.sort('age', ascending=false)

...then df_age.head(15) view 15 of people, df_age.tail(20) view 20 of youngest people.

python pandas dataframes